Skip to content

Verify footer tags when reading encrypted Parquet files with plaintext footers #7255

@adamreeve

Description

@adamreeve

Follow up task to #6637, which adds initial support for reading files that use Parquet modular encryption.

The Parquet format allows encrypting some or all column data while keeping footers in plaintext for compatibility with readers that don't support encryption. Readers that support encryption can still verify the integrity of the footer though, as a 28 byte nonce and GCM tag are written after the plaintext footer metadata (see https://github.com/apache/parquet-format/blob/master/Encryption.md#55-plaintext-footer-mode).

This should be supported in arrow-rs to allow readers to verify the integrity of plaintext footers.

This should probably be optional, eg. in C++ Parquet there's a FileDecryptionProperties::Builder::disable_footer_signature_verification method to allow disabling this.

Metadata

Metadata

Assignees

No one assigned

    Labels

    enhancementAny new improvement worthy of a entry in the changelogparquetChanges to the parquet crate

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions